How to Move Beyond a Monolithic Data Lake to a Distributed Data Mesh